Ask HN: How to improve data science skills as a software dev?
Hey HN, I'm a software engineer wanting to become good in data science. I'm already good enough at the math I think - linear algebra, analysis, diff eqs, probability, etc. - aren't a problem. I know theoretically it should be easy, heck I know how to build software and I'm pretty darn good at maths.
The problem is every time I try to start learning data science I use the same approach I do for software dev and it doesn't seem to carry over.
For instance, if I want to learn about something new in the software world say about random number generators, I'll go look at how one's built in some language and then build my own (or vice versa), learning concepts I need to as I go. Or I can also just build things I need like cli tools, parsers, rest clients, etc. in the new framework/language so I don't feel like I'm wasting my time because I can use these in my work.
With data science it seems like most "learning projects" I can think of and which are applicable in my work boil down to just data wrangling and simple regressions, which after you've done a couple really don't teach you anything new.
How can I find data science projects that are more complicated (and ideally still applicable)?
Also, I'd really be interested to know what the "core" concepts a data scientist has to know are, not math skills or tooling, just the rest :) ?
I've also been thinking about offering to do data science work to anyone for free just to get exposure to more complicated problems. Ideally, I could find a real data scientist who could just shove stuff my way and whom I could learn from. Does this make sense?
Any links or advice greatly appreciated. Thanks in advance!
This post does not have any comments yet